the structure of the proton and the...
TRANSCRIPT
The structure of the proton and the LHCMaria Ubiali
University of Cambridge
26th Rencontres de Blois, Particle Physics and Cosmology, Blois, France20th May 2014
PDFs: why bother?
PDF uncertainty dominant for many SM predictions
Precision in predicting BSM heavy particles is limited by large quark-antiquark uncertainties at large x. Reduction of theoretical uncertainty achieved by using resummed predictions still has 15% uncertainty due to PDFs
J. Campbell, ICHEP 2012 Beenakker et al (2011)
PDFs LHCPDFs uncertainties are a crucial input at the LHC, often being the limiting factor in the accuracy of theoretical predictions, both SM and BSM
Exploit the power of precise LHC data to reduce PDF uncertainties and discriminate among PDF sets
J. Campbell, ICHEP 2012 Beenakker et al (2011)
PDFs: why bother?
Outline
Introduction
Parton Distribution Functions
PDFs at the LHC
The state of the art
LHC for PDFs
Constraints from data
Conclusions and outlook
Parton Distribution Functions
PDFs cannot be computed in perturbative QCD but they are universal and their evolution with the scale is predicted by pQCD
Dokshitzer, Gribov, Lipatov, Altarelli, Parisi renormalization group equations
LO - Dokshitzer; Gribov, Lipatov; Altarelli, Parisi, 1977 NLO - Floratos,Ross,Sachrajda; Floratos,Lacaze,Kounnas, Gonzalez-Arroyo,Lopez,Yndurain; Curci,Furmanski Petronzio, 1981 NNLO - Moch, Vermaseren, Vogt, 2004
µ
2 @ f(x, µ2)
@ µ
2=
Z 1
z
dz
z
↵s
2⇡P (z) f
⇣x
z
, µ
2⌘
d�
pp!abH
dX
=
NfX
i,j=1
fi(x1, µF ) fj(x2, µF )d�
ij!abH
dX
(x1x2Shad,↵s(µR), µF )+O ⇤2nQCD
S
nhad
!
Parton Distribution Functions
They can be extracted from available experimental data and used as phenomenologicalinput for theory predictions
Different data constrain different parton combinations at different x
d�
pp!abH
dX
=
NfX
i,j=1
fi(x1, µF ) fj(x2, µF )d�
ij!abH
dX
(x1x2Shad,↵s(µR), µF )+O ⇤2nQCD
S
nhad
!
Constraints from data (pre-LHC)
DIS data
q,qbar at x > 10-4
g at small and medium x
deuteron data: disentangle isospin triplet and singlet contributions
neutrino DIS data: handle on strange
Constraints from data (pre-LHC)
Tevatron jets Good consistency with DIS data, i.e. scaling violation Largest impact on large-x gluon Significant improvements in accuracy, uncertainty reduced by factor of 2 for 0.1 < x < 0.7
gluon
R. Ball et al. Nucl. Phys. B838 (2010) 136
Constraints from data (pre-LHC)
Drell-Yanproduction
light quarks
(u - d)
Old fixed-target DY and Tevatron vector boson production data constrain light quark separation and disentangle quark-antiquark distributions
�
DY,p / u(x1)u(x2) + d(x1)d(x2)
�
DY,d / u(x1)(u+ d)(x2) + d(x1)(u+ d)(x2)
R. Ball et al. Nucl. Phys. B838 (2010) 136
_ _
PDFs: state of the art
Progress in PDF determination
< 2002: sets without uncertainty 2003-2004: first MRST, CTEQ, Alekhin sets with uncertainties 2004-now: huge progress made in statistical and theoretical understand, new players
PDG “Structure Functions”2013
Modern sets of PDFs
May 2014 CT10(w) MSTW2008 NNPDF23 ABM12 HERAPDF15
Fixed Target DIS ✔ ✔ ✔ ✔ ✘
HERA ✔ ✔ ✔ ✔ ✔
Fixed Target DY ✔ ✔ ✔ ✔ ✘
Tevatron W,Z ✔ ✔ ✔ ✘ ✘
Tevatron jets ✔ ✔ ✔ ✘ ✘
LHC data ✘ ✘ ✔ ✔ ✘
Stat. treatment Hessian∆χ²= 100
Hessian∆χ² dynamical Monte Carlo Hessian
∆χ²=1Hessian∆χ²=100
Parametrization Pol. (26 pars) Pol. (20 pars) NN (259 pars) Pol. (14 pars) Pol. (14 pars)
HQ scheme ACOT-! TR’ FONLL FFN TR’
"S Varied Fitted+varied Varied Fitted Varied
LHAPDF6.1.0 - https://lhapdf.hepforge.org
Predictions for the LHC
Gluon fusion initiated Higgs productions
Wide spread of predictions limits accuracy in Higgs characterization
Global sets quite close to each others and compatible to HERA analysis
Larger discrepancies with ABM and JR
Similar (worse) situation for ttbar cross section predictions
S. Forte, G. Watt, ArXiv:1301.6754
Parton Luminosities
G. Watt, arXiv:1301.6754
�ij(M2X) =
1
s
Z 1
⌧
dx
x
fi(x,M2X) fj
⇣⌧
x
,M
2X
⌘
Directly connected with gluon-gluon luminosity
Predictions for the LHC
For W and Z productions (quark dominated) situation is less dramatic
Predictions mostly close to each others
More significant discrepancies with ABM and JR
Compatible with data, although data the more precise the more discriminating
G. Watt, arXiv:1301.6754
Parton Luminosities
�ij(M2X) =
1
s
Z 1
⌧
dx
x
fi(x,M2X) fj
⇣⌧
x
,M
2X
⌘
G. Watt, arXiv:1301.6754
Directly connected with quark-quark luminosity
How to combine them?
➡ Envelopes [PDF4LHC prescription arXiv 1101.0538]➡ Statistical combination from different PDF groups generating MC sets. [Forte, Watt, 2013] Smaller uncertainty than envelope: 4.8% vs 3.4% for gg>H➡ Meta-PDFs: fit with input functional form the CT, MSTW and NNPDF shapes and combine in a unique consistent set [Gao, Nadolsky, 2014]➡ Crucial to decide optimal value of αS and its uncertainty in combination
G. Watt, PDF4LHC 2013
LHC data for PDFs
LHC kinematics coverage Inclusive jets and dijets (medium/large x) Isolated photon and γ+jets (medium/large x) Top pair production (large x) High pT Z(+jets) distribution (small/medium x) High pT W(+jets) ratios (medium/large x) W and Z production (medium x) Low and high mass Drell-Yan (small and large x) Wc (strangeness at medium x)
Low and high mass Drell-Yan WW production
{
{{
GLU
ON
QU
AR
KS
PHO
TON
• Jets are traditional source of information on gluon and αS
• Large-x is the region where gluons and quarks are mostly unconstrained • Wealth of precise experimental measurements• Theoretical calculation: NLO and partially NNLO gg initiated contribution has been calculated
• [Gehrmann et al]
Martin et al, (2009) ATLAS, EPJC (2013) 73 2509
GluonsInclusive Jets: data
GluonsInclusive Jets: impact of data
R. Ball et al, Nucl. Phys. B867 (2013) 244
+ ATLAS jets:90 points
Tevatron jets186 points
• NNPDF2.3 analysis includes ATLAS 35 pb-1 inclusive jet data• Moderate but significant constraint and reduction in gluon uncertainty • Many more data released: CMS full 7/8 GeV dataset, ATLAS 7/8 TeV and 2.76 TeV data• Ratio of observable at different CoM energies strongly constraint due to correlations (ATLAS)• CMS sizeable impact assessed by MSTW via reweighting
ATLAS, EPJC (2013) 73 2509
GluonsInclusive Jets: impact of data• NNPDF2.3 analysis includes ATLAS 35 pb-1 inclusive jet data• Moderate but significant constraint and reduction in gluon uncertainty • Many more data released: CMS full 7/8 GeV dataset, ATLAS 7/8 TeV and 2.76 TeV data• Ratio of observable at different CoM energies strongly constraint due to correlations (ATLAS)• CMS sizeable impact assessed by MSTW via reweighting [Ball et al (2012)]
CMS-PAS-SMP-12-012R. Thorne, QCD@LHC 2013
S. Farry, QCD@LHC 2013
• Prompt photon production directly sensitive to the gluon-quark luminosity via Compton scattering• Isolated prompt photon data well described by NLO QCD theory • ATLAS and CMS measurements at 7 TeV constrain medium-x region
D’Enterria, Rojo, Nucl. Phys. B860 (2012), 311
GluonsPrompt photon production: data
GluonsPrompt photon production: impact of data
D’Enterria, Rojo, Nucl. Phys. B860 (2012), 311
• Included ATLAS 880 nb -1, and 35 pb -1, CMS 1.9 -1 and 36 pb -1
• Moderate uncertainty reduction in the region which affects and reduce uncertainties for Higgs gluon fusion predictions by 20%• Issue: there is not yet a public fast interface available for JETPHOX [P. Aurenche et al] in PDF fits but it is likely to be available shortly
• Prompt photon production directly sensitive to the gluon-quark luminosity via Compton scattering• Isolated prompt photon data well described by NLO QCD theory • ATLAS and CMS measurements at 7 TeV constrain medium-x region
• At LHC, dominant channel is gg fusion• Exp: precise measurements of total xsec by ATLAS and CMS + differential distributions • Theory: full NNLO for total cross sections
[Czakon et al] and NLO + NNLL code for differential distributions public soon [Guzzi et al]
• Significant constraints for gluon [Czakon et al, Beneke et al]
GluonsTop pair production: data and impact
Czakon et al, JHEP 1307 (2013) 167
• In global fits, medium/large x gluon is mainly constrained by jet data. • W/Z boson at large pT (associated with jets) would provide a complementary constraint in x region which enters gg>H production• At large pT, gluon up (for Z and W+) or gluon down (for W-) scattering dominate: can exploit these observables to constrain gluon and u/d ratio
GluonsHigh pT vector boson production
MadGraph + PYTHIA
Malik,Watt ArXiv: 1304.2424
• pT spectra affected by possibly large theoretical uncertainties, soft resummation and EW corrections at small/large pT. • Need NNLO, hopefully not too far after calculation of H+j at NNLO [Boughezal et al]• Exploit ratios to cancel theoretical uncertainties
LHCb-CONF-2013-005
Quark flavor separationA wealth of data from LHC
W/Z rapidity distribution, both central (CMS and ATLAS) and forward (LHCb) Constrain quark flavor separation in a wide x range
W charge lepton asymmetry (ATLAS and CMS) Strong constraints on up and down valence quarks and sea asymmetry
CMS PRL 109 (2012)
Quark flavor separationA wealth of data from LHC
High and low mass Drell-Yan distributions provide valuable constrain to quarks and antiquarks in large and small x regions x
01,2 =
Mps
e
±Y
CMS JHEP 2013 ATLAS arXiv 1404.1212
• In global analyses strangeness is mostly determined by DIS fixed target data (CHORUS, NuTeV) -> suppressed strange sea
rs(x,Q2) =
s(x,Q2) + s(x,Q2)
2d(x,Q2)
• ATLAS analysis, based on the HERAFITTER approach, points to a non-suppressed strangeness• NNPDF2.3 analysis confirms the central value of the ATLAS analysis but finds larger uncertainties.
R. Ball et al, Nucl. Phys. B867 (2013) 244 ATLAS, Phys. Rev. Lett. 109 (2012) 012001
StrangenessA “strange” story
• W+charm data from ATLAS and CMS (both inclusive and distributions) provide a cleaner set of data to constrain strangeness from collider data• ATLAS data consistent with large s, opposite to CMS data consistent with suppressed s• Recent from NOMAD: charm dimuon production in neutrino-iron scattering consistent with NuTeV• Ultimate answer comes from inclusion of W+c data in PDF fits
StrangenessA “strange” story
R. Plakakyte DIS 2014 CMS JHEP (2014)
• EW corrections become relevant at the current precision level• Their inclusion requires PDF with QED effects• NNPDF23QED is new PDF set with uncertainties which incorporates (N)NLO QCD + LO QED effects• Photon PDF fitted from DIS and DY data (on-shell W,Z production and low/high mass DY)• Photon PDF is poorly determined from DIS data. Need hadron collider processes where photon contributes at LO!
Q = 100GeV
PhotonElectroweak corrections
Ball et al, 1308.0598
Boughezal, Liu, Petriello, ArXiv:1312.4535
Ball et al, 1308.0598
small x
large x
PhotonConstraints from LHC
• WW production is phenomenologically relevant as a background for BSM searches• At high MWW, photon-induced contribution become relevant• The large uncertainty at large MWW comes from the large uncertainty of photon PDF for x > 0.1• New LHC data give unique opportunity of constraining the photon in that region
PhotonMore constraints from LHC Ball et al, 1308.0598
Conclusions and outlookPDF uncertainties are often limiting factor in achieving precise predictionsLHC data have huge potential in constraining PDFs
Most collaborations working in inclusion of LHC data (NNPDF3.0)Towards collider only fits?Theoretical accuracy must catch up with experimental accuracy
✓ Inclusive jets and dijets both central and forward: large-x quarks and gluons ✓ Inclusive W and Z productions and asymmetries: quark-flavor separation, strangeness, photon✓ Isolated photon: medium-x gluon ✓ High-pT Z/W transverse momentum distribution✓ W+charm: direct handle on strangeness✓ W,Z + jets: gluon at medium-x and u/d ratio ✓ Off-resonance Drell-Yan at small and large mass: quarks & photon at small and large x✓ WW production: photon and light quarks✓ Top quark inclusive and differential distribution: large-x gluon ✓ Ratios at different CME✓ Z+c: intrinsic charm PDF✓ Z+b: gluon and bottom PDF ✓ Single top production: gluon and bottom PDFs